This study compared the antimicrobial resistance (AMR) of Escherichia coli detected from single samples vs. pooled samples at herd level. The national monitoring dataset included isolates from one sample per pig holding, whereas the research study included isolates from pooled samples of 10 pigs per holding. In both datasets, caecal samples were collected from healthy pigs randomly selected at slaughterhouses and plated on non-selective and antibiotic selective media. Resistance against a panel of nine antibiotics was compared between datasets by generalized linear mixed effects models (GLMMs) and by bootstrapped generalized linear model (GLM) to account for pooling. The highest proportion of resistant E. coli was observed against tetracycline and ampicillin in both datasets. In non-selective media, single and pooled samples showed similar results, but the bootstrapped GLM detected significantly lower resistance to ciprofloxacin and nalidixic acid in the national dataset. In selective media, a significantly greater proportion of resistant isolates was observed in the research dataset for ceftazidime (OR: 0.05, 95%CI = 0.01-0.42) and nalidixic acid (OR: 0.15, 95%CI = 0.05-0.51). The results suggest that one sample per holding provides similar information on AMR at herd level as pooled samples for most of the tested antibiotics, although less resistance to ciprofloxacin, ceftazidime, and nalidixic acid was detected.
Keywords: antimicrobial resistance; commensal E. coli; pigs; single/pooled samples; surveillance.
© The Author(s) 2023. Published by Oxford University Press on behalf of Applied Microbiology International.